A most challenging problem in biology is to detemine how different cell types are generated from the same genome. As a model system to understand the mechanisms involved in directed changes of cell in a eucaryote, we have exploited a system of heribale swithches of the mating type locus in the fission yeast, Schizosaccharomyces pombe. The P (plus) and M (minus) mating types of S. pombe interchange to the other cell type by a transposition-substitution reaction in which a copy of either one of two unexpressed donor loci, mat2-P and mat3-M, is transmitted to the expressed locus, mat 1. One striking observation is that within a pair of sister cells, in as many as 92% of cell divisions only one of the two cells generates a single switched cell in the next generation. We have recently shown that this pattern of switching in a cell lineage is due to an "epidgenetic" chromosomal modification such that some chromosomes are potentiated to switch whereas others are not. Furthermore, we have shown that the developmental asymmetry among daughter cells is a consequence of inheriting non-equivalent parental DNA chains. Our goal is to define the molecular nature of the "semi-heritable" DNA "modification" and the subsequent events required for switching. We will directly test the restriction-modification model we have proposed for initiating the mat 1 switching event, which is comprised of a site-specific double stranded DNA break. More specifically, we will characterize the function of the switch (swi) genes, which we have suggested to be required for the modification as well as DNA cleavage events, using both genetic and molecular techniques. The nature of the break and the sequences required for moditication and DNA cleavage will be defined. Understanding the molecular nature of the developmental program that leads to cellular differentiation in S. pombe can be expected to have general biological relevance since all ercaryotes are also expected to execute such programs.